---
title: ERNIE-Image-Turbo-MLX-4bit
canonical_url: "https://www.modelscope.cn/models/themindstudio/ERNIE-Image-Turbo-MLX-4bit"
md_url: "https://www.modelscope.cn/models/themindstudio/ERNIE-Image-Turbo-MLX-4bit.md"
repository: themindstudio/ERNIE-Image-Turbo-MLX-4bit
last_updated: 2026-06-08
license: apache-2.0
base_model:
  - baidu/ERNIE-Image-Turbo
base_model_relation: quantized
parameters: 1.9B
tensor_type:
  - BF16
  - U32
library_name:
  - safetensors
downloads: 179
stars: 0
---

# ERNIE-Image-Turbo-MLX-4bit

> ERNIE-Image-Turbo-MLX-4bit - themindstudio 在 ModelScope 开源的模型。ERNIE-Image-Turbo-MLX-4bit

themindstudio/ERNIE-Image-Turbo-MLX-4bit 是 ModelScope 魔搭社区上的 1.9B 参数机器学习模型，采用 apache-2.0 许可，基于 baidu/ERNIE-Image-Turbo 构建。

- **Repository**: themindstudio/ERNIE-Image-Turbo-MLX-4bit
- **License**: apache-2.0
- **Parameters**: 1.9B
- **Base model**: baidu/ERNIE-Image-Turbo
- **Downloads**: 179
- **Stars**: 0
- **Last updated**: 2026-06-08

Source: https://www.modelscope.cn/models/themindstudio/ERNIE-Image-Turbo-MLX-4bit

---

# ERNIE-Image-Turbo-MLX-4bit

This repository contains the 4-bit quantized version of **Baidu's ERNIE-Image-Turbo** text-to-image model, optimized for Apple Silicon using MLX.

ERNIE-Image-Turbo is a fast, high-quality 12B parameter text-to-image transformer. Quantized to 4-bit, the model size is reduced to approximately 6.5 GB, allowing it to run smoothly on Macs with 8GB or 16GB of unified memory without triggering Metal GPU watchdog timeouts.

## Model Details
- **Original Model**: [baidu/ERNIE-Image-Turbo](https://huggingface.co/baidu/ERNIE-Image-Turbo)
- **MindCraft Studio**: [MindCraft Studio Homepage](https://themindstudio.cc/mindcraft)
- **Quantization**: 4-bit weight-only quantization (`quantize=4`) via MLX.
- **Inference Steps**: Recommended 8 steps.
- **Guidance Scale**: 1.0 (CFG is disabled/forced to 1.0 for the Turbo architecture).

## How to Use in MindCraft Studio
1. Open [MindCraft Studio](https://themindstudio.cc/mindcraft).
2. Go to **Settings** -> **Model Library** -> **Import Remote Model**.
3. Point to this directory.
4. The model will show up as `ERNIE-IMAGE` (Turbo) and is ready for high-speed generation.

## Credits & References
- **Original Model**: [baidu/ERNIE-Image-Turbo](https://huggingface.co/baidu/ERNIE-Image-Turbo)
- **MindCraft Studio**: [themindstudio.cc/mindcraft](https://themindstudio.cc/mindcraft)
- **MLX Framework**: [Apple Machine Learning Research](https://github.com/ml-explore/mlx)